def overloadDir(self, here, there):
"""Copy files recursively. Overwrite local copies if they exist"""
for f in listdir(there):
fSrc = path.join(there, f)
fDst = path.join(here, f)
if path.isdir(fSrc):
if not path.exists(fDst):
self.info("Creating directory", fDst)
mkdir(fDst)
elif not path.isdir(fDst):
self.error("Destination path", fDst,
"exists, but is no directory")
self.overloadDir(fDst, fSrc)
elif path.isfile(fSrc):
isThere = False
rmDest = None
if path.exists(fDst):
isThere = True
elif path.splitext(fSrc)[1]==".gz" and \
path.exists(path.splitext(fDst)[0]):
rmDest = path.splitext(fDst)[0]
elif path.splitext(fSrc)[1]=="" and \
path.exists(fDst+".gz"):
rmDest = fDst + ".gz"
if rmDest:
remove(rmDest)
if isThere:
if not path.isfile(fDst):
self.error("Desination", fDst, "exists but is no file")
self.info("Copying", fSrc, "to", fDst)
copy(fSrc, fDst)
else:
self.errr("Source file", fSrc, "is neither file nor directory")
def clear(self):
"""Wipe the directory clean"""
for v in self.values:
nm = path.join(self.name, v)
remove(nm)
remove(nm + ".gz")
self.reread(force=True)
def clear(self):
"""Wipe the directory clean"""
for v in self.values:
nm=path.join(self.name,v)
remove(nm)
remove(nm+".gz")
self.reread(force=True)
def __remove(self,key):
f=path.join(self.name,key)
if path.exists(f):
remove(f)
elif path.exists(f+".gz"):
remove(f+".gz")
else:
error("Problem:",key,"(",f,") is supposed to exists, but no file found")
self.values.remove(key)
def __remove(self, key):
f = path.join(self.name, key)
if path.exists(f):
remove(f)
elif path.exists(f + ".gz"):
remove(f + ".gz")
else:
error("Problem:", key, "(", f,
") is supposed to exists, but no file found")
self.values.remove(key)
def run(self):
files=self.parser.getArgs()
for f in files:
if not path.exists(f):
self.error("File",f,"does not exists")
if not path.islink(f):
self.warning("File",f,"is not a symbolic link")
continue
real=path.realpath(f)
while path.islink(real) and self.opts.followSymlinks:
real=path.realpath(real)
bakName=f+".backupfileForSymlink"
copytree(real,bakName)
remove(f)
rename(bakName,f)
def cleanExtension(self, startDir, ext):
"""Go recursivly through directories and remove all files that have a specific extension"""
self.info("Looking for extension", ext, "in", startDir)
for f in listdir(startDir):
if f[0] == ".":
self.info("Skipping", f)
continue
src = path.join(startDir, f)
if path.splitext(src)[1] == ext or path.splitext(
path.splitext(src)[0])[1] == ext:
if not path.isdir(src):
self.info("Removing", src)
remove(src)
if path.isdir(src):
self.cleanExtension(src, ext)
def cleanExtension(self, startDir, ext):
"""Go recursivly through directories and remove all files that have a specific extension"""
if self.opts.verbose:
print_("Looking for extension", ext, "in", startDir)
for f in listdir(startDir):
if f[0] == ".":
if self.opts.verbose:
print_("Skipping", f)
continue
src = path.join(startDir, f)
if path.splitext(f)[1] == ext and not path.isdir(src):
if self.opts.verbose:
print_("Removing", src)
remove(src)
elif path.isdir(src):
self.cleanExtension(src, ext)
def run(self):
files = self.parser.getArgs()
for f in files:
if not path.exists(f):
self.error("File", f, "does not exists")
if not path.islink(f):
self.warning("File", f, "is not a symbolic link")
continue
real = path.realpath(f)
while path.islink(real) and self.opts.followSymlinks:
real = path.realpath(real)
bakName = f + ".backupfileForSymlink"
copytree(real, bakName)
remove(f)
rename(bakName, f)
def cleanExtension(self,
startDir,
ext):
"""Go recursivly through directories and remove all files that have a specific extension"""
self.info("Looking for extension",ext,"in",startDir)
for f in listdir(startDir):
if f[0]==".":
self.info("Skipping",f)
continue
src=path.join(startDir,f)
if path.splitext(src)[1]==ext or path.splitext(path.splitext(src)[0])[1]==ext:
if not path.isdir(src):
self.info("Removing",src)
remove(src)
if path.isdir(src):
self.cleanExtension(src,ext)
def run(self):
config=ConfigParser.ConfigParser()
files=self.parser.getArgs()
good=config.read(files)
# will work with 2.4
# if len(good)!=len(files):
# print "Problem while trying to parse files",files
# print "Only ",good," could be parsed"
# sys.exit(-1)
benchName=config.get("General","name")
if self.opts.nameAddition!=None:
benchName+="_"+self.opts.nameAddition
if self.opts.foamVersion!=None:
benchName+="_v"+self.opts.foamVersion
isParallel=config.getboolean("General","parallel")
lam=None
if isParallel:
nrCpus=config.getint("General","nProcs")
machineFile=config.get("General","machines")
if not path.exists(machineFile):
self.error("Machine file ",machineFile,"needed for parallel run")
lam=LAMMachine(machineFile,nr=nrCpus)
if lam.cpuNr()>nrCpus:
self.error("Wrong number of CPUs: ",lam.cpuNr())
print "Running parallel on",lam.cpuNr(),"CPUs"
if config.has_option("General","casesDirectory"):
casesDirectory=path.expanduser(config.get("General","casesDirectory"))
else:
casesDirectory=foamTutorials()
if not path.exists(casesDirectory):
self.error("Directory",casesDirectory,"needed with the benchmark cases is missing")
else:
print "Using cases from directory",casesDirectory
benchCases=[]
config.remove_section("General")
for sec in config.sections():
print "Reading: ",sec
skipIt=False
skipReason=""
if config.has_option(sec,"skip"):
skipIt=config.getboolean(sec,"skip")
skipReason="Switched off in file"
if self.opts.excases!=None and not skipIt:
for p in self.opts.excases:
if fnmatch(sec,p):
skipIt=True
skipReason="Switched off by pattern '"+p+"'"
if self.opts.cases!=None:
for p in self.opts.cases:
if fnmatch(sec,p):
skipIt=False
skipReason=""
if skipIt:
print "Skipping case ..... Reason:"+skipReason
continue
sol=config.get(sec,"solver")
cas=config.get(sec,"case")
pre=eval(config.get(sec,"prepare"))
preCon=[]
if config.has_option(sec,"preControlDict"):
preCon=eval(config.get(sec,"preControlDict"))
con=eval(config.get(sec,"controlDict"))
bas=config.getfloat(sec,"baseline")
wei=config.getfloat(sec,"weight")
add=[]
if config.has_option(sec,"additional"):
add=eval(config.get(sec,"additional"))
print "Adding: ", add
util=[]
if config.has_option(sec,"utilities"):
util=eval(config.get(sec,"utilities"))
print "Utilities: ", util
nr=99999
if config.has_option(sec,"nr"):
nr=eval(config.get(sec,"nr"))
sp=None
if config.has_option(sec,"blockSplit"):
sp=eval(config.get(sec,"blockSplit"))
toRm=[]
if config.has_option(sec,"filesToRemove"):
toRm=eval(config.get(sec,"filesToRemove"))
setInit=[]
if config.has_option(sec,"setInitial"):
setInit=eval(config.get(sec,"setInitial"))
parallelOK=False
if config.has_option(sec,"parallelOK"):
parallelOK=config.getboolean(sec,"parallelOK")
deMet=["metis"]
if config.has_option(sec,"decomposition"):
deMet=config.get(sec,"decomposition").split()
if deMet[0]=="metis":
pass
elif deMet[0]=="simple":
if len(deMet)<2:
deMet.append(0)
else:
deMet[1]=int(deMet[1])
else:
print "Unimplemented decomposition method",deMet[0],"switching to metis"
deMet=["metis"]
if isParallel==False or parallelOK==True:
if path.exists(path.join(casesDirectory,sol,cas)):
benchCases.append( (nr,sec,sol,cas,pre,con,preCon,bas,wei,add,util,sp,toRm,setInit,deMet) )
else:
print "Skipping",sec,"because directory",path.join(casesDirectory,sol,cas),"could not be found"
else:
print "Skipping",sec,"because not parallel"
benchCases.sort()
parallelString=""
if isParallel:
parallelString=".cpus="+str(nrCpus)
resultFile=open("Benchmark."+benchName+"."+uname()[1]+parallelString+".results","w")
totalSpeedup=0
minSpeedup=None
maxSpeedup=None
totalWeight =0
runsOK=0
currentEstimate = 1.
print "\nStart Benching\n"
csv=CSVCollection("Benchmark."+benchName+"."+uname()[1]+parallelString+".csv")
# csvHeaders=["description","solver","case","caseDir","base",
# "benchmark","machine","arch","cpus","os","version",
# "wallclocktime","cputime","cputimeuser","cputimesystem","maxmemory","cpuusage","speedup"]
for nr,description,solver,case,prepare,control,preControl,base,weight,additional,utilities,split,toRemove,setInit,decomposition in benchCases:
# control.append( ("endTime",-2000) )
print "Running Benchmark: ",description
print "Solver: ",solver
print "Case: ",case
caseName=solver+"_"+case+"_"+benchName+"."+uname()[1]+".case"
print "Short name: ",caseName
caseDir=caseName+".runDir"
csv["description"]=description
csv["solver"]=solver
csv["case"]=case
csv["caseDir"]=caseDir
csv["base"]=base
csv["benchmark"]=benchName
csv["machine"]=uname()[1]
csv["arch"]=uname()[4]
if lam==None:
csv["cpus"]=1
else:
csv["cpus"]=lam.cpuNr()
csv["os"]=uname()[0]
csv["version"]=uname()[2]
workDir=path.realpath(path.curdir)
orig=SolutionDirectory(path.join(casesDirectory,solver,case),
archive=None,
paraviewLink=False)
for a in additional+utilities:
orig.addToClone(a)
orig.cloneCase(path.join(workDir,caseDir))
if oldApp():
argv=[solver,workDir,caseDir]
else:
argv=[solver,"-case",path.join(workDir,caseDir)]
run=BasicRunner(silent=True,argv=argv,logname="BenchRunning",lam=lam)
runDir=run.getSolutionDirectory()
controlFile=ParameterFile(runDir.controlDict())
for name,value in preControl:
print "Setting parameter",name,"to",value,"in controlDict"
controlFile.replaceParameter(name,value)
for rm in toRemove:
fn=path.join(caseDir,rm)
print "Removing file",fn
remove(fn)
for field,bc,val in setInit:
print "Setting",field,"on",bc,"to",val
SolutionFile(runDir.initialDir(),field).replaceBoundary(bc,val)
oldDeltaT=controlFile.replaceParameter("deltaT",0)
for u in utilities:
print "Building utility ",u
execute("wmake 2>&1 >%s %s" % (path.join(caseDir,"BenchCompile."+u),path.join(caseDir,u)))
print "Preparing the case: "
if lam!=None:
prepare=prepare+[("decomposePar","")]
if decomposition[0]=="metis":
lam.writeMetis(SolutionDirectory(path.join(workDir,caseDir)))
elif decomposition[0]=="simple":
lam.writeSimple(SolutionDirectory(path.join(workDir,caseDir)),decomposition[1])
if split:
print "Splitting the mesh:",split
bm=BlockMesh(runDir.blockMesh())
bm.refineMesh(split)
for pre,post in prepare:
print "Doing ",pre," ...."
post=post.replace("%case%",caseDir)
if oldApp():
args=string.split("%s %s %s %s" % (pre,workDir,caseDir,post))
else:
args=string.split("%s -case %s %s" % (pre,path.join(workDir,caseDir),post))
util=BasicRunner(silent=True,argv=args,logname="BenchPrepare_"+pre)
util.start()
controlFile.replaceParameter("deltaT",oldDeltaT)
# control.append(("endTime",-1000))
for name,value in control:
print "Setting parameter",name,"to",value,"in controlDict"
controlFile.replaceParameter(name,value)
print "Starting at ",asctime(localtime(time()))
print " Baseline is %f, estimated speedup %f -> estimated end at %s " % (base,currentEstimate,asctime(localtime(time()+base/currentEstimate)))
print "Running the case ...."
run.start()
speedup=None
cpuUsage=0
speedupOut=-1
try:
speedup=base/run.run.wallTime()
cpuUsage=100.*run.run.cpuTime()/run.run.wallTime()
except ZeroDivisionError:
print "Division by Zero: ",run.run.wallTime()
if not run.runOK():
print "\nWARNING!!!!"
print "Run had a problem, not using the results. Check the log\n"
speedup=None
if speedup!=None:
speedupOut=speedup
totalSpeedup+=speedup*weight
totalWeight +=weight
runsOK+=1
if maxSpeedup==None:
maxSpeedup=speedup
elif speedup>maxSpeedup:
maxSpeedup=speedup
if minSpeedup==None:
minSpeedup=speedup
elif speedup<minSpeedup:
minSpeedup=speedup
print "Wall clock: ",run.run.wallTime()
print "Speedup: ",speedup," (Baseline: ",base,")"
print "CPU Time: ",run.run.cpuTime()
print "CPU Time User: "******"CPU Time System: ",run.run.cpuSystemTime()
print "Memory: ",run.run.usedMemory()
print "CPU Usage: %6.2f%%" % (cpuUsage)
csv["wallclocktime"]=run.run.wallTime()
csv["cputime"]=run.run.cpuTime()
csv["cputimeuser"]=run.run.cpuUserTime()
csv["cputimesystem"]=run.run.cpuSystemTime()
csv["maxmemory"]=run.run.usedMemory()
csv["cpuusage"]=cpuUsage
if speedup!=None:
csv["speedup"]=speedup
else:
csv["speedup"]="##"
csv.write()
resultFile.write("Case %s WallTime %g CPUTime %g UserTime %g SystemTime %g Memory %g MB Speedup %g\n" %(caseName,run.run.wallTime(),run.run.cpuTime(),run.run.cpuUserTime(),run.run.cpuSystemTime(),run.run.usedMemory(),speedupOut))
resultFile.flush()
if speedup!=None:
currentEstimate=totalSpeedup/totalWeight
if self.opts.removeCases:
print "Clearing case",
if speedup==None:
print "not ... because it failed"
else:
print "completely"
rmtree(caseDir,ignore_errors=True)
print
print
if lam!=None:
lam.stop()
print "Total Speedup: ",currentEstimate," ( ",totalSpeedup," / ",totalWeight, " ) Range: [",minSpeedup,",",maxSpeedup,"]"
print runsOK,"of",len(benchCases),"ran OK"
resultFile.write("Total Speedup: %g\n" % (currentEstimate))
if minSpeedup and maxSpeedup:
resultFile.write("Range: [ %g , %g ]\n" % (minSpeedup,maxSpeedup))
resultFile.close()
def run(self):
config = ConfigParser.ConfigParser()
files = self.parser.getArgs()
good = config.read(files)
# will work with 2.4
# if len(good)!=len(files):
# print_("Problem while trying to parse files",files)
# print_("Only ",good," could be parsed")
# sys.exit(-1)
benchName = config.get("General", "name")
if self.opts.nameAddition != None:
benchName += "_" + self.opts.nameAddition
if self.opts.foamVersion != None:
benchName += "_v" + self.opts.foamVersion
isParallel = config.getboolean("General", "parallel")
lam = None
if isParallel:
nrCpus = config.getint("General", "nProcs")
machineFile = config.get("General", "machines")
if not path.exists(machineFile):
self.error("Machine file ", machineFile,
"needed for parallel run")
lam = LAMMachine(machineFile, nr=nrCpus)
if lam.cpuNr() > nrCpus:
self.error("Wrong number of CPUs: ", lam.cpuNr())
print_("Running parallel on", lam.cpuNr(), "CPUs")
if config.has_option("General", "casesDirectory"):
casesDirectory = path.expanduser(
config.get("General", "casesDirectory"))
else:
casesDirectory = foamTutorials()
if not path.exists(casesDirectory):
self.error("Directory", casesDirectory,
"needed with the benchmark cases is missing")
else:
print_("Using cases from directory", casesDirectory)
benchCases = []
config.remove_section("General")
for sec in config.sections():
print_("Reading: ", sec)
skipIt = False
skipReason = ""
if config.has_option(sec, "skip"):
skipIt = config.getboolean(sec, "skip")
skipReason = "Switched off in file"
if self.opts.excases != None and not skipIt:
for p in self.opts.excases:
if fnmatch(sec, p):
skipIt = True
skipReason = "Switched off by pattern '" + p + "'"
if self.opts.cases != None:
for p in self.opts.cases:
if fnmatch(sec, p):
skipIt = False
skipReason = ""
if skipIt:
print_("Skipping case ..... Reason:" + skipReason)
continue
sol = config.get(sec, "solver")
cas = config.get(sec, "case")
pre = eval(config.get(sec, "prepare"))
preCon = []
if config.has_option(sec, "preControlDict"):
preCon = eval(config.get(sec, "preControlDict"))
con = eval(config.get(sec, "controlDict"))
bas = config.getfloat(sec, "baseline")
wei = config.getfloat(sec, "weight")
add = []
if config.has_option(sec, "additional"):
add = eval(config.get(sec, "additional"))
print_("Adding: ", add)
util = []
if config.has_option(sec, "utilities"):
util = eval(config.get(sec, "utilities"))
print_("Utilities: ", util)
nr = 99999
if config.has_option(sec, "nr"):
nr = eval(config.get(sec, "nr"))
sp = None
if config.has_option(sec, "blockSplit"):
sp = eval(config.get(sec, "blockSplit"))
toRm = []
if config.has_option(sec, "filesToRemove"):
toRm = eval(config.get(sec, "filesToRemove"))
setInit = []
if config.has_option(sec, "setInitial"):
setInit = eval(config.get(sec, "setInitial"))
parallelOK = False
if config.has_option(sec, "parallelOK"):
parallelOK = config.getboolean(sec, "parallelOK")
deMet = ["metis"]
if config.has_option(sec, "decomposition"):
deMet = config.get(sec, "decomposition").split()
if deMet[0] == "metis":
pass
elif deMet[0] == "simple":
if len(deMet) < 2:
deMet.append(0)
else:
deMet[1] = int(deMet[1])
else:
print_("Unimplemented decomposition method", deMet[0],
"switching to metis")
deMet = ["metis"]
if isParallel == False or parallelOK == True:
if path.exists(path.join(casesDirectory, sol, cas)):
benchCases.append(
(nr, sec, sol, cas, pre, con, preCon, bas, wei, add,
util, sp, toRm, setInit, deMet))
else:
print_("Skipping", sec, "because directory",
path.join(casesDirectory, sol, cas),
"could not be found")
else:
print_("Skipping", sec, "because not parallel")
benchCases.sort()
parallelString = ""
if isParallel:
parallelString = ".cpus=" + str(nrCpus)
resultFile = open(
"Benchmark." + benchName + "." + uname()[1] + parallelString +
".results", "w")
totalSpeedup = 0
minSpeedup = None
maxSpeedup = None
totalWeight = 0
runsOK = 0
currentEstimate = 1.
print_("\nStart Benching\n")
csv = CSVCollection("Benchmark." + benchName + "." + uname()[1] +
parallelString + ".csv")
# csvHeaders=["description","solver","case","caseDir","base",
# "benchmark","machine","arch","cpus","os","version",
# "wallclocktime","cputime","cputimeuser","cputimesystem","maxmemory","cpuusage","speedup"]
for nr, description, solver, case, prepare, control, preControl, base, weight, additional, utilities, split, toRemove, setInit, decomposition in benchCases:
# control.append( ("endTime",-2000) )
print_("Running Benchmark: ", description)
print_("Solver: ", solver)
print_("Case: ", case)
caseName = solver + "_" + case + "_" + benchName + "." + uname(
)[1] + ".case"
print_("Short name: ", caseName)
caseDir = caseName + ".runDir"
csv["description"] = description
csv["solver"] = solver
csv["case"] = case
csv["caseDir"] = caseDir
csv["base"] = base
csv["benchmark"] = benchName
csv["machine"] = uname()[1]
csv["arch"] = uname()[4]
if lam == None:
csv["cpus"] = 1
else:
csv["cpus"] = lam.cpuNr()
csv["os"] = uname()[0]
csv["version"] = uname()[2]
workDir = path.realpath(path.curdir)
orig = SolutionDirectory(path.join(casesDirectory, solver, case),
archive=None,
paraviewLink=False)
for a in additional + utilities:
orig.addToClone(a)
orig.cloneCase(path.join(workDir, caseDir))
if oldApp():
argv = [solver, workDir, caseDir]
else:
argv = [solver, "-case", path.join(workDir, caseDir)]
run = BasicRunner(silent=True,
argv=argv,
logname="BenchRunning",
lam=lam)
runDir = run.getSolutionDirectory()
controlFile = ParameterFile(runDir.controlDict())
for name, value in preControl:
print_("Setting parameter", name, "to", value,
"in controlDict")
controlFile.replaceParameter(name, value)
for rm in toRemove:
fn = path.join(caseDir, rm)
print_("Removing file", fn)
remove(fn)
for field, bc, val in setInit:
print_("Setting", field, "on", bc, "to", val)
SolutionFile(runDir.initialDir(),
field).replaceBoundary(bc, val)
oldDeltaT = controlFile.replaceParameter("deltaT", 0)
for u in utilities:
print_("Building utility ", u)
execute("wmake 2>&1 >%s %s" % (path.join(
caseDir, "BenchCompile." + u), path.join(caseDir, u)))
print_("Preparing the case: ")
if lam != None:
prepare = prepare + [("decomposePar", "")]
if decomposition[0] == "metis":
lam.writeMetis(
SolutionDirectory(path.join(workDir, caseDir)))
elif decomposition[0] == "simple":
lam.writeSimple(
SolutionDirectory(path.join(workDir, caseDir)),
decomposition[1])
if split:
print_("Splitting the mesh:", split)
bm = BlockMesh(runDir.blockMesh())
bm.refineMesh(split)
for pre, post in prepare:
print_("Doing ", pre, " ....")
post = post.replace("%case%", caseDir)
if oldApp():
args = string.split("%s %s %s %s" %
(pre, workDir, caseDir, post))
else:
args = string.split(
"%s -case %s %s" %
(pre, path.join(workDir, caseDir), post))
util = BasicRunner(silent=True,
argv=args,
logname="BenchPrepare_" + pre)
util.start()
controlFile.replaceParameter("deltaT", oldDeltaT)
# control.append(("endTime",-1000))
for name, value in control:
print_("Setting parameter", name, "to", value,
"in controlDict")
controlFile.replaceParameter(name, value)
print_("Starting at ", asctime(localtime(time())))
print_(
" Baseline is %f, estimated speedup %f -> estimated end at %s "
% (base, currentEstimate,
asctime(localtime(time() + base / currentEstimate))))
print_("Running the case ....")
run.start()
speedup = None
cpuUsage = 0
speedupOut = -1
try:
speedup = base / run.run.wallTime()
cpuUsage = 100. * run.run.cpuTime() / run.run.wallTime()
except ZeroDivisionError:
print_("Division by Zero: ", run.run.wallTime())
if not run.runOK():
print_("\nWARNING!!!!")
print_(
"Run had a problem, not using the results. Check the log\n"
)
speedup = None
if speedup != None:
speedupOut = speedup
totalSpeedup += speedup * weight
totalWeight += weight
runsOK += 1
if maxSpeedup == None:
maxSpeedup = speedup
elif speedup > maxSpeedup:
maxSpeedup = speedup
if minSpeedup == None:
minSpeedup = speedup
elif speedup < minSpeedup:
minSpeedup = speedup
print_("Wall clock: ", run.run.wallTime())
print_("Speedup: ", speedup, " (Baseline: ", base, ")")
print_("CPU Time: ", run.run.cpuTime())
print_("CPU Time User: "******"CPU Time System: ", run.run.cpuSystemTime())
print_("Memory: ", run.run.usedMemory())
print_("CPU Usage: %6.2f%%" % (cpuUsage))
csv["wallclocktime"] = run.run.wallTime()
csv["cputime"] = run.run.cpuTime()
csv["cputimeuser"] = run.run.cpuUserTime()
csv["cputimesystem"] = run.run.cpuSystemTime()
csv["maxmemory"] = run.run.usedMemory()
csv["cpuusage"] = cpuUsage
if speedup != None:
csv["speedup"] = speedup
else:
csv["speedup"] = "##"
csv.write()
resultFile.write(
"Case %s WallTime %g CPUTime %g UserTime %g SystemTime %g Memory %g MB Speedup %g\n"
% (caseName, run.run.wallTime(), run.run.cpuTime(),
run.run.cpuUserTime(), run.run.cpuSystemTime(),
run.run.usedMemory(), speedupOut))
resultFile.flush()
if speedup != None:
currentEstimate = totalSpeedup / totalWeight
if self.opts.removeCases:
print_("Clearing case", end=" ")
if speedup == None:
print_("not ... because it failed")
else:
print_("completely")
rmtree(caseDir, ignore_errors=True)
print_()
print_()